Early Cardiac Dysfunction in Adolescents With Low Vigorous Physical Activity

Authors: Samuel J. Burden, Najmeh Zare, Oliver J. Rider, Helen Dawes, Alexander Jones

Abstract

Background

Adverse left ventricular diastolic function (LVDF) is an early marker of cardiac dysfunction that worsens with age and can lead to heart failure. It is unclear when this deterioration begins and whether physical activity (PA) influences it. We assessed the independent relationships of adverse LVDF in adolescents with different PA intensities, compared to the association with adiposity. The impact of adverse LVDF from low PA on cardiorespiratory fitness was examined.

Methods

In 127 adolescents (aged 11‐18 years), we assessed LVDF by echocardiography, 7‐day PA by wrist‐worn accelerometry, adiposity Z scores by bioimpedance, and peak oxygen consumption by cardiopulmonary exercise testing (OxSOCRATES [Oxfordshire Sedentariness, Obesity, and Cardiometabolic Risk in Adolescents: A Trial of Exercise in Schools] study; NCT04118543). Adverse LVDF was defined using a body surface area–adjusted septal early diastolic tissue peak velocity Z score (<−2) based on pediatric reference values.

Results

Most participants with overweight (96%) or obesity (90%) had low vigorous PA (VPA). Rates of adverse LVDF were higher in groups with normal weight/low VPA (22%), overweight (28%), and obesity (45%) compared with the normal weight/high VPA group (0%; all P<0.001). Greater VPA Z score reduced the odds of adverse LVDF (odds ratio [OR], 0.20 [0.10–0.40]), independent (P=0.016) of the weaker relationship with adiposity Z score (OR, 0.57 [0.40–0.82]). Less intense PA was not associated with LVDF. Causal mediation analysis showed that those who do 5 versus 15 minutes of daily VPA had 2.4 mL/kg per minute lower peak oxygen consumption because of adverse LVDF (indirect effect).

Conclusions

Low VPA in adolescence, independent of adiposity, is linked to adverse LVDF. Such impairment predicts worse peak oxygen consumption. VPA appears important for preserving LVDF in the young.

Research PerspectiveWhat Is New? It is unclear whether low physical activity (PA) is associated with adverse diastolic function in adolescents, whether the intensity of PA matters, and whether adverse left ventricular diastolic function from low PA impacts cardiorespiratory fitness.A low duration of daily vigorous PA, but not of lesser intensities, was linked to adverse left ventricular diastolic function independently of and more strongly than adiposity status, and this impairment predicted worse cardiorespiratory fitness. What Question Should Be Addressed Next? A randomized controlled trial of vigorous PA is needed to confirm our findings and determine whether increasing vigorous PA improves diastolic function, resulting in improved cardiorespiratory fitness.

The increasing prevalence of cardiovascular disease (CVD) risk factors such as obesity and physical inactivity have contributed to a rise in heart failure mortality rates, particularly in those aged <45 years. ^1^ The deterioration of left ventricular diastolic function (LVDF) is a hallmark feature of this, ^2^ , ^3^ , ^4^ , ^5^ and longitudinal data show that most people with impaired LVDF in early‐to‐middle adulthood continue to have it as they age and that it typically worsens. ^2^ , ^3^ More recently, latent class analysis of 31 echocardiography parameters found that CVD risk was highest in those with a pattern of suboptimal diastolic function rather than suboptimal systolic function, independent of traditional CVD markers. ^4^ In support of this, lower cardiorespiratory fitness in young adulthood was associated with worse diastolic but not systolic function 25 years later, independently of adult cumulative CVD risk factor burden. ^6^ Similarly, large population studies have shown that the risk of heart failure with preserved ejection fraction (HFpEF), of which impaired LVDF is the primary pathological feature, increases progressively with higher levels of obesity and lower levels of physical activity (PA) and cardiorespiratory fitness, although no such associations were observed for heart failure with reduced ejection fraction. ^7^ , ^8^ , ^9^ Animal studies of volume overload and poor cardiometabolic health provide mechanistic insights into this, showing that the myocardium typically remodels to preserve ejection fraction at the expense of impaired LVDF. ^10^ , ^11^ Taken together, these studies provide evidence that impaired LVDF is one of the earliest signs of cardiac dysfunction, and so identification of when and why such impairment first appears should offer the best opportunity for the primary prevention of CVD.

LVDF impairment is already identifiable in young people with obesity, ^12^ hypertension, ^13^ and type 2 diabetes, ^14^ and is expected to become more prevalent given the rise of CVD risk factors in children and adolescents. ^15^ , ^16^ , ^17^ , ^18^ , ^19^ Although PA is an established risk factor for adult HFpEF, ^7^ it has not been established whether PA is associated with an early cardiac dysfunction in adolescents. PA is a readily modifiable risk factor that has well‐established and wide‐ranging cardioprotective benefits, including reduced inflammation, improved hemodynamics, better vascular endothelial function, improved autonomic balance, better lipid and thrombotic profiles, improved insulin sensitivity, and better glucose homeostasis. ^20^ Although weight loss is a potential result of PA, these benefits do not always depend on weight loss. ^21^ , ^22^ In a large longitudinal cohort, adults who were physically active children were less likely to have adverse LVDF. ^23^ This finding was independent of their adult blood pressure, PA levels, and adiposity, suggesting that childhood PA had lasting cardiovascular benefits. However, because LVDF was not assessed in childhood, it is not clear if the adult benefits reflected persistence of improved childhood LVDF or some other benefit of PA. It is also not clear whether PA intensity matters or whether PA benefits are due to weight loss or some other effect. Exercise programs with or without dieting in adolescents with obesity improve measures of LVDF, ^24^ , ^25^ , ^26^ , ^27^ , ^28^ but these studies have not specifically addressed PA intensity or the effects in those with normal weight. Furthermore, the impact of adverse LVDF on intermediate clinical outcome measures such as cardiorespiratory fitness has not been investigated. Vigorous PA (VPA) has been shown to be the only PA intensity to independently influence cardiorespiratory fitness in adolescents and to have the greatest influence on a range of cardiovascular risk factors, ^29^ , ^30^ , ^31^ , ^32^ , ^33^ but its importance as a determinant of LVDF is unknown.

The principal aim of this study was to assess the independent relationships of different intensities of routine PA to early cardiac dysfunction (abnormal LVDF) in adolescents, defined using pediatric reference data, and compare these to the better‐studied relationship with adiposity. Secondarily, we aimed to determine whether the extent of such adverse LVDF in childhood might already have a meaningful impact on exercise capacity, which is an important clinical feature in adult heart failure. ^34^ , ^35^

METHODS

Study Design

A cross‐sectional study of deidentified data collected during the baseline assessments of the OxSOCRATES (Oxfordshire Sedentariness, Obesity, and Cardiometabolic Risk in Adolescents: A Trial of Exercise in Schools; NCT04118543) was undertaken. The study was approved by the University of Oxford Ethics Committee (reference R54302/RE006) and was in accordance with the Declaration of Helsinki. Data are available upon reasonable request to the corresponding author.

Participants

Participants aged 11 to 18 years were recruited through secondary schools and academic institutions in Oxfordshire, United Kingdom, following participation in a school‐based physical education prescreening session, as described previously, ^29^ or by poster, media, or email advertisements. Participants who were willing and able to register their informed assent (aged <16 years) or informed consent (aged ≥16 years) to participation and whose parent(s)/guardian(s), where appropriate, gave informed consent for participation of their child in the study were recruited. Participants were excluded if they had contraindications for exercise determined by the Physical Activity Readiness Questionnaire; safety issues due to behavioral/intellectual limitations, or medical conditions such as neuromuscular disorders, uncontrolled epilepsy, or congenital heart disease; dairy allergy or type 1 diabetes (due to ingestion of a dairy‐based high‐calorie drink for the OxSOCRATES protocol); or contraindications for magnetic resonance imaging. Participants with normal weight and those with overweight or obesity were recruited. Normal weight participants were preferentially recruited if an earlier prescreening assessment of PA showed that they were in the upper or lower quartile of VPA at the time of that assessment, based on the larger data set derived from the prescreening study. ^29^ This was done to maximize the potential to determine relationships of PA with cardiovascular health, with a focus on VPA being due to earlier evidence that this intensity has the greatest independent effect on cardiovascular health in the young. ^29^ , ^30^ , ^31^ , ^32^ , ^33^

Anthropometrics

Height to the nearest 0.1 cm and weight to the nearest 0.1 kg were measured using a portable Harpenden Stadiometer (Holtain, Crymych, UK) and a SECA medical 770 digital floor scale (SECA, Hamberg, Germany), respectively, following standard practice. Age‐ and sex‐appropriate body mass index Z scores were calculated using World Health Organization standards. ^36^ Overweight/obesity was defined as a body mass index Z score >1 (overweight, >1–≤2; obesity, >2) and normal weight as a body mass index Z score ≥−2 to ≤1. Self‐reported Tanner score was used for the assessment of puberty stage. ^37^

Adiposity

Bioelectrical impedance analysis was assessed using a Tanita MC‐780 body composition monitor (Tanita, Tokyo, Japan). Total fat mass (kilograms) from the Tanita MC‐780 was recorded, which is comparable to dual‐energy x‐ray absorptiometry (r=0.93, P<0.001; intraclass correlation coefficient=0.88 [95% CI, 0.84–0.91]) and is highly reproducible over 3 consecutive days in adolescents (test–retest intraclass correlation coefficient=0.99). ^38^ The fat‐mass index (FMI) was calculated by dividing fat mass (kilograms) by height (meters) squared. Age‐ and sex‐specific FMI Z scores were calculated based on normative values of adolescents with normal weight from a comparable population. ^39^

Physical Activity

Adolescents were provided with a small, waterproof, triaxial, wrist‐worn accelerometer (AX3; Axivity, Melton Park, United Kingdom) and were asked to wear this continuously for 7 consecutive days. Devices were set up and processed according to a standard protocol (see Supplement for details). ^29^ Briefly, accelerations up to 16 times the force of gravity were sampled at 100 Hz and processed into bandpass‐filtered followed by Euclidian norm values, a standard metric that compares well to gold‐standard measures of routine PA in children and adolescents. ^40^ Routine PA at different intensities (sedentary time ≤1.5 metabolic equivalents [METs]; light >1.5 and ≤3.0 METs; moderate >3.0 and ≤6 METs; and vigorous >6.0 METs) ^41^ was derived from bandpass‐filtered followed by Euclidian norm data as detailed in the Supplement, using a standard approach. Guidelines suggest that the following activities are good examples of VPA where energy is routinely found to be >6.0 METs: running, swimming, fast bike riding, and basketball. Stationary (no movement) and sleep periods were identified and excluded so that only activities when participants were awake were processed.

Participants were categorized as low (≤14 minutes per day) or high (>14 minutes per day) VPA based on the previous findings that cardiovascular risk reduction might require adolescents to participate in more VPA, ^29^ , ^30^ , ^31^ , ^32^ , ^33^ and that 14 minutes of VPA was associated with median cardiorespiratory fitness in the OxSOCRATES population. ^29^ Sex‐adjusted VPA Z scores were calculated from our larger, unselected OxSOCRATES prescreening data set to determine normality in this population (n=735).

Cardiorespiratory Fitness

Participants completed a breath‐by‐breath analysis cardiopulmonary exercise test (CPET) on an ergometer bike (Lode Corival CPET, Groningen, the Netherlands) using the Godfrey protocol (see Supplement for detail). ^42^ Oxygen consumption (VO2) was assessed using a Cortex MetaLyzer 3B (Cortex Biophysik, Leipzig, Germany). Those who did not complete a maximal test were excluded from analyses. We report results on peak VO2 per kilogram (milliliters per kilogram per minute) as per the recent clinical consensus statement, ^34^ but we also used VO2 peak per kilogram of fat‐free mass (FFM) (milliliters per kilogram FFM per minute) for comparison, because FFM is known to be associated with VO2 peak. ^43^ VO2 peak Z scores were calculated using adolescent reference values from a large, contemporary population with similar characteristics to our study population and where reference values had been specifically designed to address the confounding effect of extreme weight distributions. ^44^

Echocardiography

Two‐dimensional echocardiography was completed by a consultant pediatric cardiologist using a General Electric Vivid‐i portable ultrasound system with a 3S‐RS transducer (GE Healthcare, Chicago, IL). Up to 10 cardiac cycles for each LVDF measure were recorded and averaged to account for respiratory variation. Echocardiograms were analyzed offline using Horos (Horosproject.org).

LVDF parameters were measured according to international guidelines, using standardized techniques. ^45^ Pulsed‐wave Doppler at the tips of the mitral valve, aligned with mitral inflow, was used to measure peak early and late mitral inflow velocities (E wave and A wave, respectively) and the E/A ratio was calculated. Continuous‐wave Doppler was placed ~1 cm into the left ventricular outflow tract to simultaneously display aortic ejection and mitral inflow to measure isovolumic relaxation time. Tissue Doppler imaging (TDI) at the septal and lateral basal regions of the left ventricular myocardium was used to measure septal, lateral, and averaged peak early and late myocardial velocities (e′ and a′, respectively).

Of all LVDF measures, septal TDI measures have been shown to best detect the early adverse consequences of childhood obesity, ^12^ and a septal e′ threshold appears to best distinguish childhood cardiomyopathy from healthy controls. ^46^ Given the sensitivity of these measures, we used body surface area‐adjusted septal e′ Z scores based on normal pediatric reference values to define adverse LVDF. ^47^ A threshold of −2 Z scores (2 SD below the mean) was used as a widely accepted definition of abnormality.

Blood Pressure

The average of several supine systolic and diastolic blood pressure readings were taken during the OxSOCRATES magnetic resonance imaging protocol using a Vicorder system (80 Beats Medical, Berlin, Germany).

Statistical Analysis

All statistical analyses were completed using Stata (version 18; StataCorp, College Station, TX). Histogram plots were used to assess the normality of variables. Any nonparametrically distributed data were 0‐skewed log‐transformed by taking the natural log of the nonparametric variable after a correction factor had been applied.

Differences between groups with normal weight and high VPA, with normal weight and low VPA, with overweight, and with obesity were assessed by 1‐way ANOVA with Bonferroni correction for multiple comparisons. To assess any differences in baseline characteristics between boys and girls with and without missing data, t tests and χ^2^ tests were used. Differences in CPET outcomes between participants with adverse or normal LVDF were assessed by independent t tests. Data were summarized as mean±SD or median (interquartile range), depending on distributions.

Because LVDF was associated with age, sex, and progression through the stages of puberty (Tanner score), ^37^ these potential confounders were included as independent variables in all models. To overcome the problem of strong correlations between different PA intensities, we used partial regression modeling to calculate residualized PA intensities (the amount of PA at each intensity that is independent of activity at all other intensities; see Supplement for details). The associations of each PA intensity with LVDF were tested in separate linear regression models. Multiple regression including PA and FMI Z scores was done to control for the well‐known effect of adiposity on LVDF. ^12^ , ^48^ Heart rate was included in separate models as a potential confounder, given its known influence on LVDF. Logistic regression was used to calculate odds ratios (ORs) with robust standard errors for adverse LVDF for any PA intensities that were significantly related with LVDF in the above models, controlling for FMI Z score.

Causal mediation analysis was used to determine whether low VPA leads to adverse LVDF that leads to worse VO2 peak. This was done using the potential‐outcomes framework, which assesses the assumptions necessary for determining the causal effect of a mediator, which is not always possible with traditional mediation techniques. ^49^ See the Supplement Material for more details. Statistical significance was set at P<0.05.

RESULTS

Of the 127 participants recruited, 108 had a complete data set (demographics, FMI, PA, and LVDF); 10 participants had missing FMI data due to COVID‐19 restrictions, 4 declined the Tanner pubertal stage assessment, and 5 had invalid PA data. Maximal CPET tests were completed in n=112 of the 127 recruited (n=8 were missing due to COVID‐19 restrictions and n=7 were excluded because of a submaximal test). Baseline characteristics of participants with incomplete data did not differ significantly from those with complete data sets (Table S1). Participants’ characteristics are reported in Table 1. Of those with valid PA (n=122), n=50 had normal weight and high VPA (reference group), n=27 had normal weight and low VPA, n=25 were classified with overweight, and n=20 were classified with obesity. Importantly, n=24 out of 25 (96%) of the group with overweight and n=18/20 (90%) of the group with obesity also had low VPA. Age, sex, Tanner score, height, and blood pressure did not differ significantly across groups. The group with normal weight and high VPA did more moderate PA and light PA and were less sedentary than the groups with normal weight/low VPA, overweight, or obesity.

Left Ventricular Diastolic Function

Compared with the reference group, the groups with normal weight/low VPA, overweight, and obesity all had worse LVDF (Table 2, Table S2) and a greater likelihood of being identified as having adverse LVDF (normal weight/low VPA, n=6/27 [22%], P=0.001; overweight, n=7/25 [28%], P<0.001; obesity, n=9/20 [45%], P<0.001), with none of the reference group having adverse LVDF (Figure 1). When classified solely on VPA or weight status, LVDF was worse in those with low versus high VPA or in those with obesity versus normal weight, with the greatest differences apparent when septal LVDF measures were used (Tables S3 and S4). Notably, we found our threshold for adverse LVDF to be essentially the same as the threshold that had been found by others to best separate adolescents with cardiomyopathy from those without (Figure S1). ^46^

Association of Low PA With Adverse LVDF

Linear regression models showed that more routine VPA was associated with better LVDF, independent of adiposity, with the strongest relationship being with the septal e′/a′ ratio (r=0.44 [95% CI, 0.26–0.63], P<0.001; Table S5). An additional interaction term for VPA and adiposity in this model was not significant. Moderate and light PA, and sedentary behavior were not associated with LVDF after VPA was accounted for (Table S6). Inclusion of heart rate as a potential confounder did not meaningfully alter the results (Table S5).

Each additional minute of VPA per day decreased the odds of having adverse LVDF by 22% (OR, 0.78 [95% CI, 0.71–0.87]; P<0.001). This is equivalent to each additional minute of VPA per day increasing septal e′ by 0.06 cm/s (95% CI, 0.02–0.10, P=0.006) and septal e′ Z score by 0.024 (95% CI, 0.005–0.042, P=0.013). Similarly, each unit increase in VPA Z score, which equates to 7.7 minutes in boys and 5.8 minutes in girls, decreased the odds of having adverse LVDF by 80% (Figure 2). This was not meaningfully altered when adiposity was controlled for and the effect size of 1 SD greater VPA was significantly greater than the effect of 1 SD lower FMI (Figure 2).

Adverse LVDF and Cardiopulmonary Exercise Test Outcomes

Participants with adverse LVDF had an 8.7 lower mean VO2 peak (95% CI, −4.4 to −13.0 mL/kg per minute; P < 0.001) (Figure 3), which is equivalent to 2.5 METs lower VO2 peak (1 MET is typically reported as clinically relevant). Results persisted when we repeated analyses separately for boys and girls (Figure 3). Results were similar when we used VO2 peak per kilogram of FFM (−6.4 mL/kg^FFM^ per minute [95% CI, −2.6 to −10.1], P=0.001) or VO2 peak Z score (−0.77 Z score [95% CI, −0.24 to −1.30], P=0.005) as the dependent variable. Importantly, peak heart rate was the same across groups (−1.0 bpm [95% CI, −6.3 to 4.4], P=0.73), suggesting that each group achieved a similar peak intensity during CPET. These results persisted when we adjusted for adiposity (FMI Z score), age, and sex (−3.6 mL/kg per minute [or 1 MET] [95% CI, −6.6 to −0.6], P=0.018). Formal interaction testing showed that the sexes did not differ in the relationship between LVDF and VO2 peak (P=0.52). Similar results were obtained when using body mass index Z score. Thus, the link between LVDF and VO2 peak did not depend on adiposity.

Causal mediation analysis was done to determine whether adverse LVDF mediated the effect of VPA on VO2 peak. Compared with adolescents who did 15 minutes of daily VPA, those who did less had progressively worse VO2 peak due to adverse LVDF (indirect effect; Figure 4A), whereas those who did more demonstrated limited further improvements that seemed to plateau. The indirect effect as a percentage of the total effect was calculated at each timepoint (Figure 4). These were 33.8% at 5 minutes, 28.3% at 10 minutes, 14.9% at 20 minutes, and 10.5% at 25 minutes (Table S7). Thus, at low VPA levels, adverse LVDF contributed about a third of the total effect of VPA on VO2 peak, with the remainder coming from other effects that did not depend on LVDF (direct effect, Figure 4B). Similar results were seen when we used septal e′ or septal e′/a′ (because this was most strongly correlated with VPA) as continuous mediator variables rather than the binary septal e′ threshold variable (Figures S2 and S3), and when we instead used VO2 peak per kilogram FFM as the outcome.

DISCUSSION

In this study, we show for the first time that undertaking more daily VPA is associated with better LVDF in adolescence. Importantly, no additional, independent benefit could be demonstrated for lesser intensities of PA. Our results also suggest that the associations of VPA and of adiposity with the risk of adverse LVDF are mutually independent and additive. Furthermore, the association of low VPA with the risk of adverse LVDF was stronger than the association of adiposity with adverse LVDF. This suggests that a public health focus on weight normalization alone could be counterproductive in the prevention of LVDF impairment and eventual HFpEF. ^7^ Failure to acknowledge the importance of VPA risks demotivating those with overweight or obesity who exercise, but do not lose weight, and falsely reassuring those who achieve normal weight without exercise.

We used a widely accepted definition of abnormality (<−2 Z scores) to define adverse LVDF. Strikingly, this produced a threshold that was essentially the same as one previously shown to effectively separate similarly aged adolescents with cardiomyopathy from those without (specificity 95.7%; Figure S1), ^46^ illustrating the clinical magnitude of this definition. It follows that those defined as having adverse LVDF in our study might be considered to have a clinically significant abnormality. We found no adverse LVDF in normal‐weight adolescents undertaking more than the median amounts of VPA. By contrast, nearly a quarter of normal‐weight adolescents doing less VPA and around half of those with obesity had adverse LVDF. To determine whether such LVDF impairment might be functionally important in adolescence, we examined the relationship with cardiorespiratory fitness and found that adolescents with adverse LVDF had 2.5 METs lower VO2 peak than those without such impairment. Causal mediation analysis was used to show that a substantial portion of the link between routine VPA and cardiorespiratory fitness in our study population could be explained by LVDF. Thus, there may already be meaningful functional limitations in adolescents as a result of their adverse LVDF that have not been previously recognized.

Importance of Early Prevention of Later Disease

Although we do not know whether adolescents with adverse LVDF in our study might someday be at greater risk of developing HFpEF, several strands of evidence suggest that preserving LVDF in the young may be important. Impaired LVDF is the primary pathophysiological abnormality in HFpEF. Once identified in adulthood, LVDF impairment tends to track into later life, worsening with age, ^2^ , ^3^ and is a strong risk factor for incident heart failure and death. ^4^ , ^5^ Physical inactivity in childhood is associated with worse LVDF in young adulthood, ^23^ potentially setting those adults on this adverse pathway. We studied adolescents aged 11 to 18 years, whose findings are likely to be relevant for such young adult populations. Childhood and adolescence may therefore offer a window of opportunity to prevent this pathology and reduce the risk of HFpEF in later life. It is uncertain at what age the benefits of exercise might diminish, if at all, but we speculate that there may be a point where the pathological changes become irreversible. We found that even in adolescence, adverse LVDF may limit VO2 peak, thus limiting individual capacity to undertake the VPA that might have beneficial effects on LVDF (negative feedback). Therefore, our work supports a life course approach to CVD prevention that might begin by focusing specifically on VPA in childhood, but longitudinal studies will be needed to confirm this.

The Case for VPA

This is the first study to address the mutually independent associations of different PA intensities with LVDF. We found that, when all intensities were accounted for, only VPA (activity >6.0 METs) was independently associated with LVDF. Every additional minute of VPA per day decreased the likelihood of adverse LVDF by 22%, and this was not meaningfully altered when we controlled for adiposity. Current guidelines recommend that adolescents should undertake at least 60 minutes of moderate to vigorous activity per day. ^41^ This daily duration can be difficult to achieve, and moderate to vigorous activity is attainable by moderate activity alone. Our findings suggest that a much shorter daily duration of VPA is likely to yield greater benefits for cardiovascular health than lesser PA intensities, supporting a growing body of literature in this field. ^29^ , ^30^ , ^31^ , ^32^ , ^33^ A focus on shorter durations of VPA in guidelines could offer simpler messaging and greater benefit. Although a comparison of the effects of moderate‐intensity continuous training and high‐intensity interval training on LVDF has been done, ^28^ it provided limited insight into the relative benefits of VPA compared with lesser PA intensities, because the average exercise heart rates for each training program suggested that both included VPA. The ideal daily duration of VPA is not clear. Longitudinal studies in adults consistently report that a relatively short duration of VPA per week is needed for protective effects, whereas at least 15 to 20 times more moderate PA is needed to achieve similar results. ^50^ , ^51^ A longer duration of regular moderate PA might be difficult to maintain for some and could explain why most adolescents fail to achieve the current recommendations. ^52^ Therefore, as proposed in an earlier study on cardiorespiratory fitness and supported by others, ^29^ , ^30^ , ^33^ a shorter, more specific public health target of ≈15 to 20 minutes of VPA alone might be easier to schedule daily. ^34^ , ^35^

Mechanisms

How VPA might improve LVDF remains to be determined. Low VPA is associated with higher levels of adiposity in children, ^32^ and excess adiposity, mainly visceral adiposity, is mechanistically linked with adverse LVDF in adults by impaired myocardial energetics, increased myocardial triglyceride content, and greater left ventricular concentric remodeling. ^53^ , ^54^ High VPA, but not lesser intensities, is associated with lower visceral adiposity in children, ^32^ and there is evidence that these adiposity‐related mechanisms of adverse LVDF originate in adolescent obesity. ^55^ However, we found the effect of VPA to be independent from adiposity. The exact mechanisms for this are not well understood, but there is evidence that crossing the anaerobic threshold during exercise triggers upregulation of many more potentially beneficial genetic pathways than aerobic activity does. ^56^ Thus, because moderate PA is primarily an aerobic activity, but VPA includes both aerobic and anaerobic activity, VPA may be needed to trigger the beneficial genetic and metabolic changes that improve LVDF. ^56^ , ^57^ Potential changes include increased myocardial glucose oxidation, decreased fatty acid oxidation, improved calcium handling, upregulation of genes involved in myocardial metabolism, and decreased myocardial stiffness (Figure 5). ^57^ , ^58^ , ^59^ , ^60^ , ^61^ , ^62^ , ^63^

Impact on Cardiorespiratory Fitness

We showed alterations of exercise physiology in those with adverse LVDF, independent of the effect of adiposity, suggesting adverse LVDF may already have functional implications and could represent the precursor of the symptoms of reduced exercise capacity typically seen in adults with HFpEF. ^34^ , ^35^ Because the adverse LVDF group had 2.5 METs lower VO2 peak, it is likely that such differences will have a noticeable functional impact during PA. ^69^ , ^70^ Although there are several potential explanations for this, our causal mediation analysis showed that adverse LVDF is a significant mediator and contributes to worse VO2 peak (Figure 5). Exercise‐stress echocardiography studies support this finding, showing a significant impact of LVDF on ventricular filling during exercise in higher classes of obesity, ^64^ and there is a strong inverse correlation between septal E/e′ ratios and VO2 peak in pediatric cardiomyopathy. ^65^ Additionally, following a 6‐month exercise intervention in adults with hypertension, improvements in LVDF were associated with better VO2 peak. The strongest relationship was with the change in e′/a′ ratio, ^66^ supporting our causal mediation findings.

Septal Diastolic Function

We found that septal TDI measures showed the greatest group differences in LVDF, aligning with our previous meta‐analysis in adolescents with obesity ^12^ and with studies in childhood cardiomyopathy ^46^ and type 1 diabetes. ^71^ The interventricular septum has been shown to remodel first in childhood obesity, ^72^ early hypertension, ^73^ and hypertrophic cardiomyopathy, ^74^ and that susceptibility to remodeling could explain the differences found in our study. Any future pediatric LVDF guidelines might consider a focus on septal myocardial velocities to identify early adverse function.

Strengths and Limitations

We defined adverse LVDF using pediatric reference values for TDI assessment in the portion of the myocardium (interventricular septum) known to remodel first. Interestingly, our threshold for this and our chosen TDI measures were the same as those previously shown to best separate adolescents with and without cardiomyopathy (Figure S1), a condition where diastolic dysfunction is a well‐recognized determinant of morbidity and mortality. ^46^ A potential link between objectively measured PA and LVDF in adolescents has not been extensively studied before. Two prior studies found no association, but this could be explained by methodological limitations in those studies (long epoch sampling rates and not accounting for the intercorrelation between PA intensities). ^33^ , ^75^ We addressed those limitations and used a more robust approach to PA assessment in children and adolescents (see Supplement Material for more detail). We used causal mediation analysis to support interpretation of the cross‐sectional associations found in this study. This approach is less susceptible to violations of model assumptions than other forms of mediation analysis, but well‐conducted exercise trials will be needed to corroborate our findings. We studied an unselected population of adolescents from collaborating schools in Oxfordshire, United Kingdom. In previous similar studies, ^29^ , ^76^ we showed that such adolescents were broadly representative of the social class and racial and ethnic makeup of the wider UK population. Replication of our findings in other populations will be necessary to demonstrate global applicability of our results.

CONCLUSIONS

Low VPA in adolescence is associated with adverse LVDF, independently of and to a greater extent than excess adiposity, and even at this age has a substantial impact on cardiorespiratory fitness. We suggest that early prevention of CVD should not focus solely on obesity but should also highlight the importance of regular VPA as a specific form of PA likely to yield the greatest benefits.

Sources of Funding

The work in this study was supported by a British Heart Foundation Intermediate Clinical Research Fellowship held by A.J. (FS/18/22/33479). N.Z. is funded by the British Heart Foundation grant held by A.J. S.J.B. was supported by the Professor Nigel Groome Studentship scheme (Oxford Brookes University) and is currently funded as part of a British Heart Foundation Special Project held at King's College London (SP/F/21/150013). O.J.R. is funded by a British Heart Foundation Senior Clinical Fellowship (FS/SCRF/22/32014). The views expressed are those of the authors and not necessarily those of the British Heart Foundation, the UK National Institute of Health Research, or the UK Department of Health.

Disclosures

None.

Supporting information